livekit-plugins-talklabs 2.0.0

TalkLabs TTS and STT plugin for LiveKit Agents

livekit-plugins-talklabs

TalkLabs TTS plugin for LiveKit Agents. Provides high-quality Brazilian Portuguese text-to-speech synthesis with streaming support.

Installation

pip install livekit-plugins-talklabs

Compatibility

This plugin is compatible with LiveKit Agents Framework and can be used as a drop-in TTS provider alongside other LiveKit plugins:

• ✅ Works with LiveKit Voice Assistant
• ✅ Compatible with all LiveKit STT providers (Deepgram, OpenAI, etc.)
• ✅ Compatible with all LiveKit LLM providers (OpenAI, Anthropic, etc.)

Usage

Basic Example

from livekit.plugins.talklabs import TalkLabsTTS

# Initialize TTS (same pattern as other LiveKit plugins)
tts = TalkLabsTTS(
    api_key="your-api-key"  # Get from https://talklabs.com.br
)

# Use with default voice (adam_rocha)
tts = TalkLabsTTS(api_key="your-api-key")

# Or specify custom settings
tts = TalkLabsTTS(
    api_key="your-api-key",
    voice="maria_silva",    # Female voice
    language="pt",          # Portuguese
    speed=1.2,             # Slightly faster
    sample_rate=24000      # High quality
)

With LiveKit Voice Assistant

from livekit.agents import JobContext, WorkerOptions, cli
from livekit.agents.voice_assistant import VoiceAssistant
from livekit.plugins.openai import LLM
from livekit.plugins.deepgram import STT
from livekit.plugins.silero import VAD
from livekit.plugins.talklabs import TalkLabsTTS
import os

async def entrypoint(ctx: JobContext):
    # Create a voice assistant with TalkLabs TTS
    assistant = VoiceAssistant(
        vad=VAD.load(),
        stt=STT(),
        llm=LLM(),
        tts=TalkLabsTTS(api_key=os.environ["TALKLABS_API_KEY"])  # Same as other plugins
    )

    await assistant.start(ctx.room)

if __name__ == "__main__":
    cli.run_app(WorkerOptions(entrypoint_fnc=entrypoint))

Complete Integration Example

from livekit.agents import JobContext, WorkerOptions, cli
from livekit.agents.voice_assistant import VoiceAssistant
from livekit.plugins.openai import LLM
from livekit.plugins.deepgram import STT
from livekit.plugins.silero import VAD
from livekit.plugins.talklabs import TalkLabsTTS
import os

async def entrypoint(ctx: JobContext):
    """LiveKit Voice Assistant with Brazilian Portuguese support."""

    # Initialize plugins - all follow same pattern
    vad = VAD.load()
    stt = STT(language="pt-BR")  # Portuguese speech recognition
    llm = LLM(model="gpt-4")
    tts = TalkLabsTTS(api_key=os.environ["TALKLABS_API_KEY"])  # Portuguese TTS

    # Create voice assistant
    assistant = VoiceAssistant(
        vad=vad,
        stt=stt,
        llm=llm,
        tts=tts,  # TalkLabs as TTS provider
    )

    # Set Portuguese context
    assistant.llm_context.append({
        "role": "system",
        "content": "You are a helpful assistant. Respond in Portuguese (pt-BR)."
    })

    await assistant.start(ctx.room)

    # Greet in Portuguese
    await assistant.say("Olá! Como posso ajudá-lo hoje?")

if __name__ == "__main__":
    cli.run_app(WorkerOptions(entrypoint_fnc=entrypoint))

Available Voices

Voice	Gender	Description
adam_rocha	Male	Default voice, natural Brazilian accent
maria_silva	Female	Clear and professional
carlos_santos	Male	Deep and authoritative
ana_costa	Female	Young and friendly

Configuration

Environment Variables

Set your API key as an environment variable:

export TALKLABS_API_KEY="your-api-key"

Then use it in your code:

tts = TalkLabsTTS(api_key=os.environ["TALKLABS_API_KEY"])

Parameters

Parameter	Type	Default	Description
api_key	str	Required	Your TalkLabs API key
voice	str	"adam_rocha"	Voice identifier
language	str	"pt"	Language code
speed	float	1.0	Speech speed (0.5-2.0)
sample_rate	int	24000	Audio sample rate
base_url	str	"https://api.talklabs.com.br"	API endpoint

Comparison with Other LiveKit TTS Plugins

Plugin	Languages	Streaming	Best For
livekit-plugins-talklabs	Portuguese (pt-BR)	✅	Brazilian Portuguese applications
livekit-plugins-openai	Multiple	✅	General purpose, English
livekit-plugins-elevenlabs	Multiple	✅	High quality English voices
livekit-plugins-cartesia	English	✅	Low latency English

Working Example

import asyncio
import os
from livekit.agents import JobContext, WorkerOptions, cli
from livekit.agents.voice_assistant import VoiceAssistant
from livekit.plugins.openai import LLM
from livekit.plugins.deepgram import STT
from livekit.plugins.silero import VAD
from livekit.plugins.talklabs import TalkLabsTTS

async def entrypoint(ctx: JobContext):
    """Minimal working example."""

    # Simple one-line TTS initialization like other plugins
    assistant = VoiceAssistant(
        vad=VAD.load(),
        stt=STT(),
        llm=LLM(),
        tts=TalkLabsTTS(api_key=os.environ["TALKLABS_API_KEY"])
    )

    await assistant.start(ctx.room)

if __name__ == "__main__":
    cli.run_app(WorkerOptions(entrypoint_fnc=entrypoint))

Features

• 🎯 Native LiveKit Integration: Works seamlessly with LiveKit Agents Framework
• 🔄 Streaming Support: Real-time audio streaming for low latency
• 🇧🇷 Brazilian Portuguese: Optimized for pt-BR pronunciation
• ⚡ Low Latency: < 200ms to first byte
• 🎵 High Quality: 24kHz sample rate
• 🔌 Plugin Architecture: Drop-in replacement for any LiveKit TTS provider

Requirements

• Python 3.9+
• LiveKit Agents 0.8.0+
• TalkLabs API key (get from talklabs.com.br)

Support

• 📧 Email: [e-mail]
• 📖 Documentation: docs.talklabs.com.br
• 🐛 Issues: GitHub Issues

License

MIT License - see LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

---

Made with ❤️ by TalkLabs for the LiveKit community